How can we better protect a transplanted kidney?
image: How can we better protect a transplanted kidney?
Credit: Wroclaw Medical University
How can we better protect a transplanted kidney?
The combination of immunology and artificial intelligence
When we think of kidney transplantation, we usually focus on surgery, immunosuppressive drugs, and early post-operative tests. Much less attention is paid to what happens months or years later, when a silent damaging process may begin inside the transplanted organ.
Under the microscope, a pathologist may detect changes indicating a specific form of rejection called microvascular inflammation (MVI)—a subtype of antibody-mediated rejection that until recently was underestimated. Diagnosing it without biopsy remains very difficult. Even when standard blood tests, such as anti-HLA antibodies, are negative, a damaging process may still be occurring in the graft.
This diagnostic gap inspired a research team from Wroclaw Medical University, led by Prof. Mirosław Banasik, to investigate additional risk factors—especially non-HLA antibodies—and to apply AI-based analytical tools. The study, conducted by Jakub Mizera, MD, was published in Frontiers in Immunology.
At the University Clinical Hospital in Wroclaw, more than 1,000 kidney transplant recipients are monitored long-term. Some have lived with their grafts for 20–30 years. The goal is to understand why certain kidneys fail despite treatment—and how to prevent premature loss.
Diagnostic criteria are changing
A transplanted kidney is structurally complex. Damage in the glomeruli or capillaries is classified as microvascular inflammation, which harms the filtration structures of the organ. Since the 2022 update of the Banff Classification, MVI is considered a serious warning sign of possible antibody-mediated rejection.
At the same time, evidence shows that antibody-mediated rejection accounts for over 50% of kidney graft losses. Yet classic diagnostics often fail: pathologists may see features of rejection in biopsy even when anti-HLA antibodies are absent.
This discrepancy encouraged researchers to explore non-HLA antibodies, particularly those targeting the angiotensin II type 1 receptor (AT1R). Because the kidney experiences unavoidable ischemic injury during transplantation, molecules such as AT1R can become exposed, potentially triggering an immune response.
Transplant archive and digital magnifying glasses
The study included 167 patients who had previously undergone biopsy following graft dysfunction. Using Banff 2022 criteria, researchers assessed each sample for MVI while also measuring AT1R antibodies in blood.
Two groups were formed:
-
79 patients with MVI,
-
88 patients without MVI.
Nearly 50% of MVI patients had positive AT1R antibodies, compared with 24% without MVI—suggesting a strong association.
The team then applied AI-based analytics—not generative AI, but statistical modeling, ROC analysis, and association rule mining—to determine at what antibody level the risk becomes clinically meaningful.
Results showed that the highest quintile of AT1R (>12 U/ml) carried a distinctly elevated risk:
-
MVI occurred in almost two-thirds of patients,
-
the risk was 37% higher than in the general study group.
Thus, not all positive results are equally meaningful—only high AT1R titers appear to be particularly dangerous.
Ongoing research now includes antibody subclasses and additional non-HLA antibodies (ETAR, ETBR, PAR1, PAR2). The long-term vision is to create a comprehensive, AI-supported non-invasive immunological profile to detect early rejection risk without biopsy.
AI as the new stethoscope of transplantology
AI will not replace clinicians, but it can help them detect patterns hidden in complex datasets. Identifying new immunological risk factors may enable earlier intervention and extend graft lifespan—reducing the need for retransplantation and improving organ availability for others.
As Prof. Banasik emphasizes, the goal is to go beyond average survival times and ensure that more patients can benefit from long-lasting graft function. AI tools, including automated biopsy analysis and advanced data modeling, will likely become standard in transplant medicine.
Young science
The project is also a success story of young research talent. MD Jakub Mizera highlights the value of interdisciplinary collaboration and the supportive research environment at Wroclaw Medical University.
Despite technical language—“microvascular inflammation,” “non-HLA antibodies”—the core question is simple: how can we help a donated kidney function for as long as possible? Identifying hidden factors contributing to graft loss is essential, and the work from Wrocław shows that combining clinical expertise with artificial intelligence may bring this goal within reach.
READ MORE: How can we better protect a transplanted kidney?
This material is based on the article:
Jakub Mizera, Piotr Donizy, Agnieszka Hałoń, Dariusz Jańczak, Marta Kępinska, Maciej Pondel, Mirosław Banasik,